Re: [RFC v4 7/8] hisi_acc_vfio_pci: Add support for VFIO live migration

[Jason Gunthorpe <jgg@nvidia.com>]

On Tue, Feb 08, 2022 at 01:34:24PM +0000, Shameer Kolothum wrote:

Overall this looks like a fine implementation, as far as I can tell it
meets the uAPI design perfectly.

Why did you decide not to do the P2P support?

> +static struct file *
> +hisi_acc_vf_set_device_state(struct hisi_acc_vf_core_device *hisi_acc_vdev,
> +			     u32 new)
> +{
> +	u32 cur = hisi_acc_vdev->mig_state;
> +	int ret;
> +
> +	if (cur == VFIO_DEVICE_STATE_RUNNING && new == VFIO_DEVICE_STATE_STOP) {
> +		ret = hisi_acc_vf_stop_device(hisi_acc_vdev);
> +		if (ret)
> +			return ERR_PTR(ret);

Be mindful that qemu doesn't handle a failure here very well, I'm not
sure we will be able to fix this in the short term.

> +static int hisi_acc_vfio_pci_init(struct hisi_acc_vf_core_device *hisi_acc_vdev)
> +{
> +	struct vfio_pci_core_device *vdev = &hisi_acc_vdev->core_device;
> +	struct pci_dev *vf_dev = vdev->pdev;
> +	struct hisi_qm *vf_qm = &hisi_acc_vdev->vf_qm;
> +
> +	/*
> +	 * ACC VF dev BAR2 region consists of both functional register space
> +	 * and migration control register space. For migration to work, we
> +	 * need access to both. Hence, we map the entire BAR2 region here.
> +	 * But from a security point of view, we restrict access to the
> +	 * migration control space from Guest(Please see mmap/ioctl/read/write
> +	 * override functions).
> +	 *
> +	 * Also the HiSilicon ACC VF devices supported by this driver on
> +	 * HiSilicon hardware platforms are integrated end point devices
> +	 * and has no capability to perform PCIe P2P.
> +	 */
> +	vf_qm->io_base =
> +		ioremap(pci_resource_start(vf_dev, VFIO_PCI_BAR2_REGION_INDEX),
> +			pci_resource_len(vf_dev, VFIO_PCI_BAR2_REGION_INDEX));
> +	if (!vf_qm->io_base)
> +		return -EIO;
> +
> +	vf_qm->fun_type = QM_HW_VF;
> +	vf_qm->pdev = vf_dev;
> +	mutex_init(&vf_qm->mailbox_lock);

mailbox_lock seems unused

> +	hisi_acc_vdev->vf_id = PCI_FUNC(vf_dev->devfn);

Does this need to use the pci_iov_vf_id() function? funcs don't need
to be tightly packed, necessarily.

This should be set when the structure is allocated, not at open time.

> +	hisi_acc_vdev->vf_dev = vf_dev;
> +	vf_qm->dev_name = hisi_acc_vdev->pf_qm->dev_name;

Also unused

> +	hisi_acc_vdev->mig_state = VFIO_DEVICE_STATE_RUNNING;
> +
> +	return 0;
> +}
>  
>  static int hisi_acc_pci_rw_access_check(struct vfio_device *core_vdev,
>  					size_t count, loff_t *ppos,
> @@ -129,63 +1067,96 @@ static long hisi_acc_vfio_pci_ioctl(struct vfio_device *core_vdev, unsigned int
>  
>  static int hisi_acc_vfio_pci_open_device(struct vfio_device *core_vdev)
>  {
> -	struct vfio_pci_core_device *vdev =
> -		container_of(core_vdev, struct vfio_pci_core_device, vdev);
> +	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(core_vdev,
> +			struct hisi_acc_vf_core_device, core_device.vdev);
> +	struct vfio_pci_core_device *vdev = &hisi_acc_vdev->core_device;
>  	int ret;
>  
>  	ret = vfio_pci_core_enable(vdev);
>  	if (ret)
>  		return ret;
>  
> -	vfio_pci_core_finish_enable(vdev);
> +	if (!hisi_acc_vdev->migration_support) {

This should just test the core flag and get rid of migration_support:

		hisi_acc_vdev->core_device.vdev.migration_flags =
			VFIO_MIGRATION_STOP_COPY;

> +++ b/drivers/vfio/pci/hisi_acc_vfio_pci.h
> @@ -0,0 +1,119 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/* Copyright (c) 2021 HiSilicon Ltd. */
> +
> +#ifndef HISI_ACC_VFIO_PCI_H
> +#define HISI_ACC_VFIO_PCI_H
> +
> +#include <linux/hisi_acc_qm.h>
> +
> +#define VDM_OFFSET(x) offsetof(struct vfio_device_migration_info, x)
> +
> +#define HISI_ACC_MIG_REGION_DATA_OFFSET                \
> +	(sizeof(struct vfio_device_migration_info))
> +
> +#define HISI_ACC_MIG_REGION_DATA_SIZE (sizeof(struct acc_vf_data))

These three are not used any more

> +struct acc_vf_data {
> +#define QM_MATCH_SIZE 32L
> +	/* QM match information */
> +	u32 qp_num;
> +	u32 dev_id;
> +	u32 que_iso_cfg;
> +	u32 qp_base;
> +	/* QM reserved 4 match information */
> +	u32 qm_rsv_state[4];
> +
> +	/* QM RW regs */
> +	u32 aeq_int_mask;
> +	u32 eq_int_mask;
> +	u32 ifc_int_source;
> +	u32 ifc_int_mask;
> +	u32 ifc_int_set;
> +	u32 page_size;
> +
> +	/* QM_EQC_DW has 7 regs */
> +	u32 qm_eqc_dw[7];
> +
> +	/* QM_AEQC_DW has 7 regs */
> +	u32 qm_aeqc_dw[7];
> +
> +	/* QM reserved 5 regs */
> +	u32 qm_rsv_regs[5];
> +
> +	/* qm memory init information */
> +	dma_addr_t eqe_dma;
> +	dma_addr_t aeqe_dma;
> +	dma_addr_t sqc_dma;
> +	dma_addr_t cqc_dma;

You can't put dma_addr_t in a structure that needs to go
on-the-wire. This should be u64

> +};
> +
> +struct hisi_acc_vf_migration_file {
> +	struct file *filp;
> +	struct mutex lock;
> +	bool disabled;
> +
> +	struct acc_vf_data vf_data;
> +	size_t total_length;
> +};
> +
> +struct hisi_acc_vf_core_device {
> +	struct vfio_pci_core_device core_device;
> +	u8 migration_support:1;
> +	/* for migration state */
> +	struct mutex state_mutex;
> +	enum vfio_device_mig_state mig_state;
> +	struct pci_dev *pf_dev;
> +	struct pci_dev *vf_dev;
> +	struct hisi_qm *pf_qm;
> +	struct hisi_qm vf_qm;
> +	int vf_id;
> +
> +	struct hisi_acc_vf_migration_file *resuming_migf;
> +	struct hisi_acc_vf_migration_file *saving_migf;
> +};
> +#endif /* HISI_ACC_VFIO_PCI_H */